The Effect of GOSAT Observations on Estimates of Net CO 2 Flux in Semi-Arid Regions of the Southern Hemisphere

Transcription

1 SOLA, 2016, Vol. 12, , doi: /sola The Effect of GOSAT Observations on Estimates of Net CO 2 Flux in Semi-Arid Regions of the Southern Hemisphere Masayuki Kondo 1, Tazu Saeki 1, Hiroshi Takagi 2, Kazuhito Ichii 1, 2, and Kentaro Ishijima 1 1 Department of Environmental Geochemical Cycle Research, Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan 2 Center for Global Environmental Research, National Institute for Environmental Studies, Tsukuba, Japan Abstract Greenhouse gases Observing SATellite (GOSAT) is the operational satellite dedicated to atmospheric CO 2 observations. Assimilation of data provided by GOSAT is expected to yield reliable CO 2 fluxes in semi-arid regions because of frequent observations owing to clear skies. Here we estimated net CO 2 flux over semi-arid regions of the Southern Hemisphere using the GOSAT column averaged CO 2 (X CO2 ) and surface CO 2 measurements. Assimilation of GOSAT X CO2 indicated that semi-arid regions are integral components of recent terrestrial CO 2 uptake, accounting for 44% globally. Compared with estimates assimilated from surface measurements, estimates by GOSAT X CO2 suggest a 50% reduction in the semi-arid CO 2 uptake, amounting to 1.1 Pg C yr 1. Significant estimation differences occurred for South America and South Africa, where the GOSAT makes frequent measurements but where surface CO 2 measurements are limited. In comparison, the two estimates varied less in Australia, where more surface measurements are available. These results suggest that GOSAT X CO2 is effective at regulating excess estimates of semi-arid CO 2 uptake in regions that are less constrained by surface CO 2 measurements. To promote understanding of climate change effects in semi-arid regions, it is important to continue monitoring trends in CO 2 uptake with GOSAT. (Citation: Kondo, M., T. Saeki, H. Takagi, K. Ichii, and K. Ishijima, 2016: The effect of GOSAT observations on estimates of net CO 2 flux in semi-arid regions of the Southern Hemisphere. SOLA, 12, , doi: /sola ) 1. Introduction The terrestrial biosphere uptakes ~2.5 Pg C (petagrams of carbon) of atmospheric CO 2 annually, amounting to 25% of annual anthropogenic CO 2 emissions (Le Quéré et al. 2013). Until recently, assessment results based on both models and field observations were in agreement that boreal and temperate ecosystems in northern middle high latitudes and tropical ecosystems were the dominant contributors to net terrestrial CO 2 uptake (e.g., McGuire et al. 2009; Phillips et al. 2009; Pan et al. 2011). However, following recent effects of climate change, there is growing evidence that this may no longer hold true. Some models have suggested that the CO 2 sequestration capability of arctic tundra and boreal forest ecosystems in northern high latitudes has weakened over the last decade (Hayes et al. 2011), while satellite-remote sensing data have inferred that ecosystems in high latitudes no longer effectively uptake CO 2 under the increased temperature conditions of recent decades (Piao et al. 2014). Similarly, studies have suggested that persistent drying since 2000 has degraded CO 2 uptake in tropical regions such as the Amazon Basin (Hilker et al. 2014), with ecosystem productivity in the area having decreased extensively following severe droughts in 2005 and 2010 (Zeng et al. Corresponding author: Masayuki Kondo, Department of Environmental Geochemical Cycle Research, Japan Agency for Marine-Earth Science and Technology, , Showa-machi, Kanazawa-ku, Yokohama, , Japan. redmk92@gmail.com. 2016, the Meteorological Society of Japan. 2008; Xu et al. 2011; Parazoo et al. 2013). In contrast, it has been suggested that semi-arid ecosystems are under a substantial positive change in net CO 2 uptake, largely due to increased availability of water resources (Poulter et al. 2014; Ahlström et al. 2015). Multi-model estimates indicate that interannual variations and trends in recent terrestrial CO 2 uptake are dominated by semi-arid ecosystems (Ahlström et al. 2015). Furthermore, work using a process model simulation has suggested that semi-arid regions in the Southern Hemisphere have been particularly influenced by climate change, specifically it was found that semi-arid regions in South America, South Africa, and Australia accounted for 51% of the global net CO 2 sink in 2011 (Poulter et al. 2014). However, these recent reports about large changes in CO 2 uptake in semi-arid regions need to be interpreted with caution because of the limited availability of validation data. For example, eddy covariance measurements of net ecosystem exchange are significantly limited in semi-arid regions, particularly in the Southern Hemisphere. Satellite remote sensing provides information about vegetation activities (e.g., vegetation indices) at a global scale, but these data only serve as indirect proxies of net CO 2 flux (Piao et al. 2013). Greenhouse gases Observing SATellite (GOSAT) CO 2 inversion is a potentially useful product for evaluating recent CO 2 flux in semi-arid regions. GOSAT was launched by the Japanese Aerospace Exploration Agency (JAXA) in January 2009 and is the first operational satellite to provide global atmospheric CO 2 concentrations in the form of column-averaged dry air molar fractions (X CO2 ) (Kuze et al. 2012). Owing to frequent occurrence of clear skies, semi-arid regions are covered by more number of GOSAT X CO2 ; therefore, reliable estimates of CO 2 uptake are expected from the assimilation of the GOSAT X CO2. This study investigated recent net CO 2 flux of semi-arid regions in the Southern Hemisphere, namely South America, South Africa, and Australia, using assimilation (i.e., CO 2 inversion) based on the GOSAT X CO2. We aimed to identify the effects of X CO2 on estimates of net CO 2 flux by comparing the results with estimates from assimilation based only on surface CO 2 measurements. 2. Materials and methods 2.1 GOSAT-based flux estimates Monthly estimates of net CO 2 flux for 64 sub-continental and ocean-basin regions (42 tiles for land and 22 tiles for ocean; Fig. 1a) are publicly distributed by the National Institute for Environmental Studies (NIES) GOSAT Project as GOSAT Level 4A data product (GOSAT L4A). The regional flux estimates are obtained by optimizing a priori flux estimates such that model-predicted concentrations given by an atmospheric transport model concur with the corresponding observations (Maksyutov et al. 2013). The a priori flux dataset, atmospheric transport model, and CO 2 observation data for constraining model simulations used in this estimation approach (GOSAT inversion) are briefly explained below. The a priori flux dataset used for GOSAT inversion comprises four components: 1) daily net ecosystem exchange (NEE) predicted by the Vegetation Integrative SImulator for Trace gases

2 182 Kondo et al., Semi-Arid CO2 Flux by GOSAT Inversion Fig. 1. Maps showing (a) the number of Greenhouse gases Observing SATellite (GOSAT) XCO2 data points per each 5 5 grid cell over the analyzed period (41 months from Jun 2009 Oct 2012). The overlaid red circles indicate the locations of CO2 measurement sites (GLOBALVIEW-CO2: GV) chosen for this study (212 sites); and (b) the three semi-arid regions of the Southern Hemisphere analyzed in this study. (VISIT) (Ito 2008); 2) monthly ocean atmosphere CO2 fluxes generated from observations of surface ocean CO2 partial pressure (pco2) (Valsala and Maksyutov 2010); 3) monthly biomass burning emissions estimated from an empirical relationship between biomass burning emission (the Global Fire Emissions Database (GFED) version 3.1 (van der Werf et al. 2010)) and burnt area (GFED burnt area version 4.0 (Giglio et al. 2013)), and; 4) monthly fossil fuel CO2 emissions obtained by merging the Open source Data Inventory of Anthropogenic CO2 emissions (ODIAC) (Oda and Maksyutov 2011) and the Carbon Dioxide Information Analysis Center datasets (Andres et al. 2011). The NIES global atmospheric tracer Transport Model (NIES-TM) was used for performing forward simulations of atmospheric CO2 (Takagi et al. 2011; Maksyutov et al. 2013) Atmospheric transport was driven by a real-time operational analysis dataset of the Japan Meteorological Agency (JMA) Climate Data Assimilation System (JCDAS) (Onogi et al. 2007). The GOSAT inversion integrates satellite-based atmospheric CO2 concentrations with surface-based data. The values of XCO2 used in this study were retrieved from high-resolution spectra of reflected sunlight in shortwave infrared (SWIR) spectral wavebands collected by the Thermal And Near infrared Sensor for carbon Observation Fourier-Transform Spectrometer (TANSOFTS) onboard the GOSAT satellite (publicly available as GOSAT FTS SWIR Level 2 data product). Details on the XCO2 retrieval approach can be found in a report by Yoshida et al. (2013). The GLOBALVIEW-CO2 surface-based CO2 data (GLOBALVIEWCO2, (2013); hereafter referred to as GV) were used in conjunction with the GOSAT XCO2 values to constrain the regional surface CO2 fluxes. 2.2 Analysis Using the GOSAT L4A monthly surface CO2 flux (version 02.03), we evaluated recent net CO2 flux in semi-arid regions of the Southern Hemisphere. We specifically considered three semi-arid regions, i.e., South America, South Africa (including Namibia, Botswana, and Zimbabwe), and Australia (including New Zealand) (Fig. 1b). These semi-arid regions were classified based on the regional boundary definition used in the TransCom atmospheric tracer transport model intercomparison project (Gurney et al. 2002), except that the classification for South Africa was modified to be close to the classification used in Poulter et al. (2014). Over these three regions, more GOSAT XCO2 retrievals are found because of frequent occurrence of clear skies associated with semi-arid regions (Fig. 1a). In order to characterize the effect of the GOSAT XCO2 retrievals, net CO2 flux from the inversion constrained only by GV data (GV-only inversion) was compared with the GOSAT inversion. The study period was set to three years (June 2009 May 2012). Taking into account the seasonal cycle in the Southern Hemisphere, the annual period was set to run from June to May of the following year. To understand the effect of the GOSAT XCO2 retrievals on semi-arid net CO2 flux estimates, seasonal variations in net CO2 flux by the GOSAT inversion were evaluated in conjunction with the uncertainty reduction rate (UR), which represents the degree of contribution from the GOSAT XCO2 retrievals in constraining regional fluxes. Based on Takagi et al. (2011), UR was calculated as follows: s UR (%) = 1 GOSAT+GV 100 sgv (1) where sgv and sgv+gosat are the uncertainties in the monthly CO2 exchanges estimated from the GV dataset and from both the GV dataset and XCO2 retrievals. 3. Results Both GOSAT and GV-only inversions showed that the semiarid regions make a large contribution to global terrestrial CO2 uptake, but the level of contribution is largely different between the two inversions (Fig. 2). Over the study period, mean annual global CO2 budgets were similar between the GOSAT and GV-only inversions ( 2.52 Pg C yr 1 and 2.64 Pg C yr 1, respectively). However, CO2 uptake in semi-arid regions, as given by the GV-only inversion ( 2.20 Pg C yr 1), was nearly two times larger than that given by the GOSAT inversion ( 1.11 Pg C yr 1), resulting in that the former constitutes 84% and the latter 44% of Fig. 2. Mean annual net CO2 flux (Pg C yr 1: June 2009 May 2012) for globe and semi-arid regions in the Southern Hemisphere by the GOSAT inversion (L4A: GOSAT+GV) and ground-based CO2 measurement only inversion (GV only).

3 SOLA, 2016, Vol. 12, , doi: /sola the global CO2 uptake. Over the study period, the GOSAT and GV-only inversions relatively agree in seasonal variations of global net CO2 flux, with similar amplitudes of CO2 sink-source patterns (Fig. 3a). In semi-arid regions in the Southern Hemisphere, the two inversions exhibited similar variation patterns of net CO2 flux, but variations of the GOSAT inversion were associated with weaker CO2 sinks and stronger CO2 sources than those of the GV-only inversion (Fig. 3b). These differences in seasonal variations between the GOSAT and GV-only inversions are more clearly illustrated scatter plots (Figs. 3c and 3d). Global net CO2 fluxes of the two inversions were closely in line with the 1:1 line, with a slope of Meanwhile in semi-arid regions in the Southern Hemisphere, it was found that the overall distribution of net CO2 fluxes was shifted toward the GOSAT inversion, despite a slope was close to the 1:1 line (i.e., 1.05). Further investigation on three semi-arid regions in the Sothern Hemisphere illustrated regional differences in net CO2 flux by the GOSAT and GV-only inversions. In South America and South Africa, two inversions exhibited similar seasonal variation patterns (Figs. 4a and 4b), but as indicated by scatter plots, these distributions tended to lean toward the GOSAT inversion (Figs. 4d and 4e). Meanwhile, seasonal variations of the GOSAT and GV-only inversions for Australia were nearly identical (Fig. 4c), with a slope and intercept of the regression in line with the 1:1 line (Fig. 4f ). Corresponding to those regional differences in net CO2 fluxes, regional differences in UR were identified in the semi-arid regions. The inclusion of the GOSAT XCO2 retrievals effectively reduced uncertainty in the South American and South African estimates: 10% and 22% on average, respectively (Figs. 4a and 4b). Compared to those two regions, UR was least profound in Australia: ~4% on average (Fig. 4c). 4. Discussions This study has highlighted the effects of GOSAT XCO2 retrievals on CO2 uptake estimates for semi-arid regions in the Southern 183 Hemisphere. The significant semi-arid CO2 uptake indicated in previous work (i.e., Poulter et al. (2014)) was noted in our results obtained with the inversion assimilating the GOSAT XCO2 retrievals. For the period from June 2009 to May 2012, both the GOSAT and GV-only inversions indicated that a large proportion of global CO2 uptake is attributable to the three Southern Hemisphere semiarid regions, the value of 84% indicated by the GV-only inversion is notably high, while the estimate by the GOSAT inversion (44%) is closer to the 2011 estimate (51%) of Poulter et al. (2014). From the comparison of the GOSAT L4A and GV-only inversions, we found that use of the GOSAT XCO2 retrievals significantly affected CO2 uptake in South America and South Africa. Corresponding to these uptake reductions, large UR ( 10%) values were also noted for these regions. In previous work, CO2 inversions based on five GOSAT XCO2 retrievals processed with different algorithms indicated a similar result, i.e., large UR ( 10%) in South America and South Africa and large reductions in the CO2 sink of South America (Takagi et al. 2014). In contrast to these two regions, the CO2 budgets estimated by the GOSAT and GV-only inversions for Australia were less varied, indicating that the contribution of XCO2 retrievals was low in this region. We regard that these influences of the GOSAT XCO2 retrievals on net CO2 fluxes are robust, as indicated by relationships between UR and differences between the GOSAT and GV-only inversions (Fig. 5). In South America and South Africa, UR and the differences in two inversions shows positive linear relationships with a statistical significance (Figs. 5a and 5b: R2 = 0.42 and 0.37, respectively, with p < 0.01), while no clear relationship was identified for Australia (Fig. 5c: R2 = 0.08, p > 0.05). Regional differences in the effects of XCO2 retrievals are attributable to the availability of surface CO2 measurements. The GOSAT XCO2 retrievals were consistently available within and in the vicinity of all three semi-arid regions (Fig. 1a). However, the UR in Australia barely exceeded 5%, unlike in South America and South Africa (Fig. 3). This is because the GOSAT XCO2 retrievals were more effective in regions under-sampled by surface observation networks (i.e., GV sites) (Maksyutov et al. 2013). In Australia, besides the four surface observation sites, aircraft observations Fig. 3. Seasonal variations in net CO2 flux (Pg C yr 1) for (a) globe and (b) semi-arid regions by the GOSAT inversion (L4A: red line for estimated values and red shading for uncertainties) and ground-based CO2 measurement only inversion (GV only: black line for estimated values and grey shading for uncertainties) for the period June 2009 May Scatter plots of seasonal variations between the GOSAT and GV-only inversions for (a) globe and (b) semi-arid regions, with grey shading indicating the 95% confidence ellipse.

4 184 Kondo et al., Semi-Arid CO2 Flux by GOSAT Inversion Fig. 4. Seasonal variations in net CO2 flux (Pg C yr 1) for (a) South America and (b) South Africa, and (c) Australia by the GOSAT and GV only inversions) for the period June 2009 May The rate of uncertainty reduction (UR: %) by the inclusion of GOSAT XCO2 retrievals is shown, along with seasonal variations in net CO2 flux. Scatter plots of seasonal variations between the GOSAT and GV-only inversions for (a) South America, (b) South Africa, and (c) Australia. Except URs, figure configurations are the same as in Fig. 3. Fig. 5. Relationships between the rate of uncertainty reduction (UR: %) and differences in seasonal variations between the GOSAT and GV-only inversions for (a) South America, (b) South Africa, and (c) Australia, with grey shading indicating the 95% confidence ellipse. were available within and in the vicinity (see Fig. 1a). However, surface observations were limited in the other regions, with just two sites in South America and one site in South Africa. Therefore, the situation in Australia represents a case where the constraints by the GV data prevail over that by the GOSAT XCO2 retrievals. Besides Australia, a previous study also identified a similar situation in North America, with this region characterized by a high availability of the XCO2 retrievals but also covered by dozens of GV sites (Maksyutov et al. 2013). It should be noted that, for the study regions, the GOSAT XCO2 retrievals affected appreciably the magnitude of net CO2 flux, but not seasonal variations. This result implies that despite the difference in magnitude, both the GOSAT and GV-only inversions would show similar responses to climate anomalies. In Fig. 6, seasonal anomalies of net CO2 flux by the GOSAT and GV-only inversions are quite similar for the semi-arid regions as expected. In particular, from the later part of 2010 to the early 2011, notable positive precipitation anomalies in the semi-arid regions promoted increases in vegetation activities represented by Normalized Difference Vegetation Index (NDVI) and in CO2 uptake for both the

5 SOLA, 2016, Vol. 12, , doi: /sola observations on CO 2 flux estimates of the terrestrial carbon cycle. In particular for regions undergoing rapid change (e.g., semi-arid regions), it is important to continue monitoring trends in net CO 2 flux with GOSAT, also using observations from additional satellites dedicated to CO 2 monitoring, such as the Orbiting Carbon Observatory-2 (OCO-2). On the same token, it is important to evaluate the semi-arid CO 2 sink using inversion systems different from this study (i.e., NIES-TM). Although effects of the GOSAT X CO2 retrievals on CO 2 uptake were found similar regardless of different retrieval algorithms (Takagi et al. 2014), there are non-negligible differences among net CO 2 fluxes estimated by different inversion systems (Houwling et al. 2015). Thus, the result from this study needs to be regarded as one of potential results of GOSAT inversions and to be validated by other inversion systems in future. Lastly, we stress importance in improving the retrieval and bias correction algorithms for the X CO2 further, as performance of those algorithms directly affect results of the inversion (e.g., Basu et al. 2013). Particularly for regions constrained by only a few ground observations such as South America and South Africa, the accuracy of the X CO2 becomes a critical factor to the inversion. Likewise, it is necessary to grow a network of ground observations for the semi-arid regions in future. This is crucial not only for strengthening constraints of the inversion system, but also for validation of the X CO2 and simulated CO 2. Acknowledgements This research was supported by Environment Research and Technology Development Funds (2-1401) from the Ministry of the Environment of Japan. GOSAT L4A is publically available from the NIES GOSAT Project web site ( gateway/gateway/menupage/open.do). The authors acknowledge all members of the GOSAT Project, which is a joint effort promoted by the Japan Aerospace Exploration Agency (JAXA), the National Institute for Environmental Studies (NIES), and the Ministry of the Environment (MOE), Japan. Fig. 6. Seasonal anomalies of net CO 2 flux by the GOSAT and GV-only inversions for (a) semi-arid regions in the Southern Hemisphere, (b) South America, (c) South Africa, and (d) Australia. In addition to net CO 2 flux, seasonal anomalies of precipitation from CRU TS 3.23 (Harris et al. 2014) and of Normalized Difference Vegetation Index (NDVI) from TERRA MODIS (Huete et al. 2002) that averaged for the regions in Fig. 1b are shown. two inversions (Fig. 6a), which is largely due to the contribution from Australia as consistent with the indication by previous works (Poulter et al. 2014; Detmers et al. 2015) (Fig. 6d). However, this result does not necessarily suggest that seasonal variations are always unaffected by the GOSAT X CO2 retrievals, because, contrary to this study, the inclusion of the GOSAT X CO2 retrievals significantly enhanced anomalies of net CO 2 flux in response to increased temperature in the case of Siberia (Kondo et al. 2015). To fully characterize the effect of the GOSAT X CO2 retrievals, it is important to continue validation studies, including evaluation of the spatio-temporal effects on biases inherited in the GOSAT X CO2 on net CO 2 flux estimation. 5. Outlook It was encouraging to find that net CO 2 fluxes varied significantly in regions characterized by the strong effect of the GOSAT X CO2 retrievals (high URs). In conjunction with previous work, which also analyzed the effect of the GOSAT X CO2 retrievals (e.g., Guerlet et al. 2013; Parazoo et al. 2013; Saeki et al. 2013; Basu et al. 2014; Deng et al. 2014; Kondo et al. 2015), the result of this study facilitates our understanding of the effect of satellite Edited by: S. Morimoto References Ahlström, A., and co-authors, 2015: The dominant role of semiarid ecosystems in the trend and variability of the land CO 2 sink. Science, 348, , doi: /science.aaa1668. Andres, R. J., J. S. Gregg, L. Losey, G. Marland, and T. A. Boden, 2011: Monthly, global emissions of carbon dioxide from fossil fuel combustion. Tellus, 63B, , doi: / j x. Basu, S., M. Krol, A. Butz, C. Clerbaux, Y. Sawa, T. Machida, H. Matsueda, C. Frankenberg, O. P. Hasekamp, and I. Aben, 2014: The seasonal variation of the CO 2 flux over Tropical Asia estimated from GOSAT, CONTRAIL, and IASI. Geophys. Res. Lett., 41, , doi: /2013gl Basu, S., S. Guerlet, A. Butz, S. Houweling, O. Hasekamp, I. Aben, P. Krummel, P. Steele, R. Langenfelds, M. Torn, S. Biraud, B. Stephens, A. Andrews, and D. Worthy, 2013: Global CO 2 fluxes estimated from GOSAT retrievals of total column CO 2. Atmos. Chem. Phys., 13, , doi: /acp Deng, F., and co-authors, 2014: Inferring regional sources and sinks of atmospheric CO 2 from GOSAT X CO2 data. Atmos. Chem. Phys., 14, , doi: /acp Detmers, R. G., O. Hasekamp, I. Aben, S. Houweling, T. T. van Leeuwen, A. Butz, J. Landgraf, P. Köhler, L. Guanter, and B. Poulter, 2015: Anomalous carbon uptake in Australia as seen by GOSAT. Geophys. Res. Lett., 42, , doi: /2015GL

6 186 Kondo et al., Semi-Arid CO 2 Flux by GOSAT Inversion Giglio, L., J. T. Randerson, and G. R. van der Werf, 2013: Analysis of daily, monthly, and annual burned area using the fourth-generation global fire emissions database (GFED4). J. Geophys. Res. Biogeosci., 118, , doi: /jgrg GLOBALVIEW-CO2, 2013: Cooperative Global Atmospheric Data Integration Project, updated annually. Multi-laboratory compilation of synchronized and gap-filled atmospheric carbon dioxide records for the period (obspack_ co2_1_globalview-co 2 _2013_v1.0.4_ ). NOAA ESRL Global Monitoring Division, Boulder, Colorado, U.S.A. (Available online at OBSPACK/1002, accessed 3 April 2016). Guerlet, S., S. Basu, A. Butz, M. Krol, P. Hahne, S. Houweling, O. P. Hasekamp, and I. Aben, 2013: Reduced carbon uptake during the 2010 Northern Hemisphere summer from GOSAT. Geophys. Res. Lett., 40, , doi: / grl Gurney, K. R., and co-authors, 2002: Towards robust regional estimates of CO 2 sources and sinks using atmospheric transport models. Nature, 415, , doi: /415626a. Harris, I., P. D. Jones, T. J. Osborn, and D. H. Lister, 2014: Updated high-resolution grids of monthly climatic observations the CRU TS3.10 Dataset. Int. J. Climatol., 34, doi: /joc Hayes, D. J., A. D. McGuire, D. W. Kicklighter, K. R. Gurney, T. J. Burnside, and J. M. Melillo, 2011: Is the northern highlatitude land-based CO 2 sink weakening? Global Biogeochem. Cy., 25, GB3018, doi: /2010gb Hilker, T., A. I. Lyapustin, C. J. Tucker, F. G. Hall, R. B. Myneni, Y. Wang, J. Bi, Y. M. de Moura, and P. J. Sellers, 2014: Vegetation dynamics and rainfall sensitivity of the Amazon. Proc. Natl. Acad. Sci. U. S. A., 111, 16041, doi: /pnas Houweling, S., and co-authors, 2015: An intercomparison of inverse models for estimating sources and sinks of CO 2 using GOSAT measurements. J. Geophys. Res. Atmos., 120, doi: /2014jd Huete, A. R., K. Didan, T. Miura, E. P. Rodriguez, X. Gao, and L. G. Ferreira, 2002: Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens. Environ., 83, , doi: /s (02) Ito, A., 2008: The regional carbon budget of East Asia simulated with a terrestrial ecosystem model and validated using AsiaFlux data. Agri. Forest Meteor., 148, , doi: /j.agrformet Kondo, M., K. Ichii, H. Takagi, and M. Sasakawa, 2015: Comparison of the data-driven top-down and bottom-up global terrestrial CO 2 exchanges: GOSAT CO 2 inversion and empirical eddy flux upscaling. J. Geophys. Res. Biogeosci., 119, doi: /2014jg Kuze, A., H. Suto, K. Shiomi, M. Nakajima, and T. Hamazaki, 2012: Level 1 algorithms for TANSO on GOSAT: processing and on-orbit calibrations. Atmos. Meas. Tech., 5, , doi: /amt Le Quéré, C., and co-authors, 2013: The global carbon budget Earth System Science Data, 5, , doi: /essd McGuire, A. D., L. G. Anderson, T. R. Christensen, S. Dallimore, L. Guo, D. J. Hayes, M. Heimann, T. D. Lorenson, R. W. Macdonald, and N. Roulet, 2009: Sensitivity of the carbon cycle in the Arctic to climate change. Ecol. Monogr., 79, , doi: / Maksyutov, S., H. Takagi, V. K. Valsala, M. Saito, T. Oda, T. Saeki, D. A. Belikov, R. Saito, A. Ito, Y. Yoshida, I. Morino, O. Uchino, R. J. Andres, and T. Yokota, 2013: Regional CO 2 flux estimates for based on GOSAT and groundbased CO 2 observations. Atmos. Chem. Phys., 13, , doi: /acp Oda, T., and S. Maksyutov, 2011: A very high-resolution (1 km 1 km) global fossil fuel CO 2 emission inventory derived using a point source database and satellite observations of nighttime lights. Atmos. Chem. Phys., 11, , doi: /acp Onogi, K., and co-authors, 2007: The JRA-25 Reanalysis. J. Meteor. Soc. Japan, 85, Pan, Y., and co-authors, 2011: A large and persistent carbon sink in the world s forests. Science, 333, , doi: / science Parazoo, N. C., and co-authors, 2013: Interpreting seasonal changes in the carbon balance of southern Amazonia using measurements of XCO 2 and chlorophyll fluorescence from GOSAT. Geophys. Res. Lett., 41, , doi: / grl Phillips, O. L., and co-authors, 2009: Drought sensitivity of the amazon rainforest. Science, 323, , doi: / science Piao, S., and co-authors, 2013: Evaluation of terrestrial carbon cycle models for their response to climate variability and to CO 2 trends. Glob. Change Biol., 19, , doi: /gcb Piao, S., and co-authors, 2014: Evidence for a weakening relationship between interannual temperature variability and northern vegetation activity. Nature Commun., 5, 5018, doi: /ncomms601. Poulter, B., D. Frank, P. Ciais, R. B. Myneni, N. Andela, J. Bi, G. Broquet, J. G. Canadell, F. Chevallier, Y. Y. Liu, S. W. Running, S. Sitch, and G. R. van der Werf, 2014: Contribution of semi-arid ecosystems to interannual variability of the global carbon cycle. Nature, 509, , doi: / nature Saeki, T., S. Maksyutov, M. Saito, V. Valsala, T. Oda, R. J. Andres, D. Belikov, P. Tans, E. Dlugokencky, Y. Yoshida, I. Morin, O. Uchin, and T. Yokota, 2013: Inverse modeling of CO 2 fluxes using GOSAT data and multi-year ground-based observations. SOLA, 9, 45 50, doi: /sola Takagi, H., T. Saeki, T. Oda, M. Saito, V. Valsala, D. Belikov, R. Saito, Y. Yoshida, I. Morino, O. Uchino, R. J. Andres, T. Yokota, and S. Maksyutov, 2011: On the benefit of GOSAT observations to the estimation of regional CO 2 fluxes. SOLA, 7, , doi: /sola Takagi, H., and co-authors, 2014: Influence of differences in current GOSAT X CO2 retrievals on surface flux estimation. Geophys. Res. Lett., 41, , doi: /2013gl Valsala, V., and S. Maksyutov, 2010: Simulation and assimilation of global ocean pco 2 and air-sea CO 2 fluxes using ship observations of surface ocean pco 2 in a simplified biogeochemical offline model. Tellus, 62B, , doi: / j x. van der Werf, G. R., J. T. Randerson, L. Giglio, G. J. Collatz, M. Mu, P. S. Kasibhatla, D. C. Morton, R. S. DeFries, Y. Jin, and T. T. van Leeuwen, 2010: Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires ( ). Atmos. Chem. Phys., 10, , doi: /acp Xu, L., A. Samanta, M. H. Costa, S. Ganguly, R. R. Nemani, and R. B. Myneni, 2011: Widespread decline in greenness of Amazonian vegetation due to the 2010 drought. Geophys. Res. Lett., 38, L07402, doi: /2011gl Yoshida, Y., and co-authors, 2013: Improvement of the retrieval algorithm for GOSAT SWIR XCO 2 and XCH 4 and their validation using TCCON data. Atmos. Meas. Tech., 6, , doi: /amt Zeng, N., J.-H. Yoon, J. A. Marengo, A. Subramaniam, C. A. Nobre, A. Mariotti, and J. D. Neelin, 2008: Causes and impacts of the 2005 Amazon drought. Environ. Res. Lett., 3, , doi: / /3/1/ Manuscript received 18 April 2016, accepted 16 June 2016 SOLA: jstage. jst. go. jp/browse/sola/